Non-alcoholic fatty liver disease (NAFL or NAFLD) and its subset, non- alcoholic steatohepatitis (NASH) are increasingly recognized as common forms of liver disease.. In the absence of concomitant cellular injury, fatty liver is a benign condition that may cause elevated liver enzymes, fatigue and abdominal pain. MASH is identified by the presence of fat in the liver plus hepatocellular injury, inflammation and varying degrees of liver fibrosis. It afflicts up to 3% of adults n the United States and one third of these people may be at risk for developing cirrhosis. NASH also affects children, although its prevalence in the pediatric population is less well defined. Currently 2% of liver transplants performed in the United States are performed because of known diagnosis of NASH. Insulin resistance, with its major associated diseases of obesity and Type 2 diabetes, is emerging as a major coexisting condition. This application proposes two clinical studies to be performed in the context of a cooperative clinical research network to achieve the long-term goals of establishing the role of hyperinsulinemia in the pathogenesis of NASH and identifying rational and effect strategies to prevent and cure NASH. These goals will be addressed by specific aims of this proposal that seek to better understand the prevalence of NASH in hyperinsulinemic patients and establish whether reducing insulin levels pharmacologically improves the necroinflammatory changes associated with NASH. Two clinical studies are proposed. The first study establishes the prevalence of NASH in patients with hyprinsulinemia and imaging evidence of fatty liver. A secondary goal of the prevalence study is to establish racial differences in the risk for developing NASH because NASH may be underrepresented or underdiagnosed in African Americans. Enrollment will include adequate African Americans to allow subgroup analysis. The second proposed study is to a 48 week treatment trail of patients with NASH using the PPAR-gamma ligand rosiglitazone and, if needed to control hyperinsulinemia, metformin. Liver biopsies of patients recruited from all Clinical Centers will be compared to liver biopsies of patients treated with the standard recommendation of weight reduction. The primary endpoint will be improvement in the liver biopsy necroinflammatory score.